lychengrex / image-denoising-with-deep-cnns Goto Github PK

Halo, I want to load a model from a another checkpoint that has been previously trained in testing. how can i do that? thank you

I have tried to train seperate models specifically for denoise picture with certain noise level. Is the model capable of training at different noise levels? When testing a pic, the model can handle pictures with different noise level at same time.
I noticed that if you try to train the model from a previous check point, they have to be consistent on the noise level, otherwise, it will show the papameter conflict.

Thank you for your time.

Any suggestions on testing it in demo.ipynb for own test-images?

Hello,
Can you help me to get and display the psnr and loss value of the result image at the testing?
Thank you

The Weshmo Plot function does not refresh in real time during training

I set plot to True, but don't display the images. Can't I run directly in PyCharm

"Cannot create this experiment: "
ValueError: Cannot create this experiment: I found a checkpoint conflicting with the current setting.

any solutions?

Hi,
thank you for sharing the code. It worked well for me on the colab. But when finish training i check the checkpoint, the folder is always empty?

Thanks.

Please tell me, how we calculate accuracy of this model.

Hello! Is there's anything to obtain checkpoint.pth.tar in here? My cpu device is very very bad for training such dataset in 200 epochs
Thank you

could i cite your article？